Hello, and welcome to the demonstration of a natural language integration with UiPath. A natural language processor, or NLP is a technology that allows computers to give meaning to user inputs.
In chatbots, voice assistants and automated emails the integration of NLP provides a humanistic touch to take the user experience to the next level.
In this demo, we’ll detail how to start the implementation of NLP using UiPath. To start the process, the first step is to pick an NLP engine and train it success will depend on how well the AI is trained.
For this demo, we chose wit.ai as it is free and provides a REST API to access it. To start you’ll have to train with AI so we can recognize certain phrases. Here you can see that if I type, “How old are you?”, that NLP will recognize the word “old” and tell me that the intention of the phrase is age. You can train the AI in several aspects such as age, date, name, and order.
In this integration, we created a small bot in UiPath that we call the NLP you will send certain questions to the NLP, and it will return the intent of those questions.
Let’s dive into the demo.
This spot only has two questions to choose from. One is, “What is your name?”, and the other is, “How old are you?”.
Note that in a real scenario, you will get any question or phrase from the user directly.
A text is then sent to the NLP. Once received, we click “OK”. After confirming our selection, we click “OK” again. Our choice is then sent to the NLP engine, the REST API. What we will get back is our text and the entities that the engine found.
Here, the engine is telling you that question intent is a name. You will also give us a value for confidence. Here it is 90% confident in its outcome, meaning the engine is 90% sure that this phrase intent contains a name.
After this, we click “OK”. And we will see the entity is a name. This process might look simple, but when implemented correctly, it was very impactful. It can identify lots of words and phrases and one phrase can contain more than one entity.
You can then decide what to do with them after. In this example, I know I need to reply with the name or retrieve a name. The NLPs are as powerful as you train them to be, which is why they are considered AI.
The best examples for this are chat bots and voice assistants like Siri, Alexa and Google Home. They use this technology to understand what you’re trying to say, even if you change the words.
So let’s try another phrase. We’ll go with the question. How old are you? This one should return an age intent.
As you can see, the engine says we are looking for an age with a 93% confidence rating. The engine that tells the engine you’re looking for is age, we can combine the phrases and get both intents.
This provides a solution to a very small area, but it will help you understand how NLP integration works and how it can be used for projects.
NLP is used in many organizations. You might experience it in automated calls when a machine prompts you by saying, “How can I help you?”.
After your response, the machine is smart enough to understand and reply back with your balance. The possibilities abound with NLP and AI-driven solutions.
And this concludes our demo for the natural language processing with UiPath. For more information, visit smartbridge.com/automation.